Light Peak panned by OEM, report

9/22/2010 11:51 AM EDT

Yeah, the information in this article by Rick reinforces the hypothesis that Light Peak is yet to gain enough strength in order to overcome the popularity of USB and USB 3.0 will be the winner at least in the near future.

9/22/2010 1:14 PM EDT

Is this a major surprise? This is a major system change. Isn't it par for the course that such a shift will take some time?

9/22/2010 3:12 PM EDT

"None of the top three (HP, Dell, Acer) have made even a hint of support for Light Peak—these OEM's are well known supporters of USB 3.0," said the report.

These OEM's have invested in 3.0 and are going to throw good money after bad.. Trouble is, if they think selling old tech in a deflationary economy is going to save them then they will go the route that the US auto industry went, (read up on the hybrid tech. history). Light peak is the right direction, only a dead man walking corp. would think that 3.0 is the way to go, just my O. ;o)

9/22/2010 3:44 PM EDT

Well I for one find it shocking that Rick could find an unnamed engineer with such bold opinions.

9/22/2010 4:39 PM EDT

Does the report list the flaws? I don't see them mentioned other than cost. In fact I see a lot of positive listed.
If LP is made cost effective enough for cellphones etc. don't you think the pc guys will utilize it? LP is not a comm. protocol like USB3.0 is - it's is a comm infrastructure. USB3.0 or 1394 (or whatever next) can utilize it. Whatcha see as an alternative, un-named engineer?

9/22/2010 5:52 PM EDT

I agree with Dylan. PC makers will eventually go to something like LP, but not now when they are in the midst of a transition to USB3.

LP has to build an ecosystem of devices to make it valuable to users, costs need to come down and we could use some apps that really need 10+Gbits/s.

The report, by the way, has tons of details...but unlike news stories, you have to buy it ;-)

9/22/2010 9:04 PM EDT

The headline does not match the story.

All the story says is LightPeak will not ramp quickly - wow, gee, neither did USB 3.0 !! That's hardly news.

Comparing with USB 3.0 alone also misses the point that LightPeak can cover much larger distances.

It is certainly never going to displace USB connectors, for a long time yet, so USB 3.0 need not be too worried.

The incremental costs will dictate what is installed on mainstream PCs, but having 10G optical is appealing, and certainly will see keen early adopters.

9/22/2010 9:28 PM EDT

It’s interesting to see how everyone compares USB 3 to LP without mentioning (or knowing) that they are fundamentally different. That difference is Power supply to the device, although LP is faster then USB it can only transfer data and no power. Don’t forget that all the gadgets (flash drives, HDDs etc.) need to be powered first before they can be used.
PS: unless you dream about your next flash drive with a cord ☺

9/22/2010 11:42 PM EDT

Few points:
- unnamed engineer equals opinion some one person.
- Lightpeak is physical layer protocol, i.e. USB 3.0 X 2 can ride on it , also protocol such as PCI-e, HDMI, SATA, etc can use this as physical layer.

Rick,
Go tell your engineer to think out of the box.

9/23/2010 12:45 AM EDT

I am also an unnamed engineer with opinions looking for an anonymous interview. And just before I am overwhelmed by excited journalists, let me say the following:
Everybody (other than Mr. T in AZ) is right and correct. The way Light Peak is touted now is as a replacement for USB 3, which is just taking its baby steps.
- The current USB 2 market is OK with actual payload transfer rates of 1-300 Mbits per second.
- Where are the possible applications? Build it and they will come only works in the movies. An ecosystem is required for a technology to succeed.
- Did I mention that $2 is a lot of money in the mass market business?
- Why is Intel trying to kill USB 3 before it is sold in its first chipset? That could be a nice assignment for an invetigative reporter - oops! no such profession

9/23/2010 1:10 AM EDT

"Why is Intel trying to kill USB 3 before it is sold in its first chipset?"

It doesn't take an investigative reporter to know why Intel wants to steer the industry down the light path, they already have a replacement for LP.

http://www.trustedreviews.com/networking/news/2010/07/28/Intel-Reveals-Faster-than-Light-Peak-Technology/p1

9/24/2010 4:10 AM EDT

" "Twice the data rate for that cost just doesn’t make sense—it's taking profit margin away," said the engineer."

No one imagined expensive accelerometer/gyroscopes implemented in cell-phones.
No one imagined expensive high-res (nowadays OLED) screens implemented in cell phones.
No one imagined expensive anything - until Apple came along with the iPhone.

Who can imagine expensive innovation in the PC OEMs?

"That said, Light Peak may be used in some server applications and it could be adopted in consumer systems from Apple"
So they'll wait for Apple to do it, and then copy it?
But why will Apple want to use this? Which devices will you connect to your computer? Does Apple control any other end-devices in this eco-system that can utilize high-data rates?

Lets see how Dell and HP will do when the next iPod series will include LP connectors, and the only PC with LP enabled will be the Mac.

Why are PC OEMs going under? I can't imagine.

9/24/2010 12:56 PM EDT

This is a pity. The greatest opportunity offered by LP is that long suffering consumers (of which we are all!), could potentially be offered just one cable plug/socket type. No more Mini/Micro/Etc USB cables and plugs, just a single cable with identical ends that would serve multiple purposes, from distributing data to power. Is utopia always tomorrow?

9/24/2010 2:09 PM EDT

LP is a major system change. Usually such changes only get deployed where performance improvements are high (ideally an order of magnitude) or cost is reduced. LP doesn't offer any of the two. Too little, too late...Kris

9/24/2010 3:19 PM EDT

Oflife might be right and it would indeed be a pity... but Intel (w/Apple it seems) will bring it on. Followers to follow, with late adopters and laggards getting punished by end users in the marketplace in all likelihood.

9/24/2010 8:16 PM EDT

Years ago I remember Intel showing USB3 off with twin optical links in the
plug, to work up to 100Gb/s, the cable carrying power. I remember Intel
making that great leap into laser on silicon, which I would have imagined
greatly moved forward their time schedule and greatly brought down their
price points (referred to in the link posted before http://www.trustedreviews.com/networking/news/2010/07/28/Intel-Reveals-Faster-than-Light-Peak-Technology/p1 ). I remember a laptop
with working Light peak on it being demonstrated, was that running of a
chip like the silicon laser demonstration in the link above?

I predict that Intel is interested in moving lightpeak technology in the
silicon laser form into a on board interconnect, then on chip, eventually
maybe, one day, more and more of an optical chip. I wonder how this will
relate to future versions of the terascale processor.

About pricing, what is $2 a port off a premium system, and after a while
when and if the silicon laser technology feature comes down to consumer
market a lot less than that.

Will people prefer their system to be outdated too quickly? Having a port
that will be mainstream/dominate by end of their systems life is useful
and worth extra money on a premium system.

9/24/2010 8:17 PM EDT

Light Peak also has extreme use as a household, and small business,
networking standard that offers obscure advantages pairing to optical
telecommunications. Data on a wireless link can be recorded and decrypted
over the years and not always as desirable. 300m is also enough to
displace some cabling in larger enterprises as well.

Does anybody remember the IEEE1394 stunting USB 1.0, that lingered for
years until USB 1.1. Recently Intel has announced the next version of PCI
Express technology. I fully expect at least a USB 3.1 proposal, or
replacement, to be announced in the next twelve months that may have the
optical links. I also imagine that Intel could potentially also retain a
lot of licensing rights to the optical technology. While you may not see a
Silicon Photonics link at the mainstream anytime soon, the technology can
turn up in light peak in years to come.

Many possibilities there.

9/24/2010 8:21 PM EDT

My apologies, line return spaces come up regularly in the editor but not in the post.


W

9/27/2010 3:41 PM EDT

If I remember correctly, at IDF last year, Light Peak was demonstrated for the first time, with Intel showing demo's and prototypes this year. Why would anyone think this was going to hit the market in a major way this year?

This is the next generation technology beyond USB 3.0, if Intel wasn't demoing it now, they wouldn't be in the position to roll out products based on it in two years from now, when it will probably being starting to take off.

12/5/2010 9:30 AM EST

"None of the top three (HP, Dell, Acer) have made even a hint of support for Light Peak"

their crazy not to also push Light peak ASAP if they want consumers to actually buy their PC's

""there's nothing compelling about Light Peak" in its initial implementation, said the PC engineer."

this engineer is on crack if he's really thinks that.

"By contrast, Light Peak's promise to double USB 3.0 data rates to nearly 10 Gbits/s "won't be that significant for a lot of apps," said the engineer."

more bull, the very first protocol Any 3rd party light peak vendor would/should include is Ethernet Over Lightpeak as a generic option, so there is a Very significant for any and all protocols that right on that already , everything from SMB (small medium business) all the way to the SOHO/Home LAN/WAN markets

all the other potential use stuff is just added gravy, Ethernet Over Lightpeak all the way ;)

12/5/2010 9:34 AM EST

"You will need higher data rates than 10Gbits/s to make differences in apps like video something end users can really see," he said."


LOL, this engineer guys never heard of the Massive AVC/H.264 HD 1080P high profile x264 Encoding ? , or the fact Intel are going to include inside Every single Sandy Bridge their HD AVC Encoder/Decoder Engine and there's even an Internal Intel x264 patch being written Right Now by Francois Piednoel , Senior Performance analyst at Intel Corp Santa Clara ... to take advantage of that internal Encoder Core ASIC, all in all many LAN connected PC's and all their app's will find a very good use for this Lightpeak 10Gig throughput




those Lazy ethernet OEM's have sat arould For Years not even trying to make and sell better than 1gig for Far to Many Year's, end user's people want to pay for this 10 gig lightpeak LAN/WAN ASAP so let them.

12/5/2010 9:35 AM EST

its just a shame we cant buy them RIGHT NOW, as i/most end users have lots of Christmas holidays money/credit burning a hope right now just waiting to be spent on something worth while like light peak LAN Kits, all the other potential use stuff is just added gravy, Ethernet Over Lightpeak all the way ;)

12/5/2010 9:52 AM EST

"I would rather use USB 3.0 and drive three meters which is more than adequate for any usage model out there," said the engineer."

LOL, sure he would, everyone else would rather have their up to 100 meter/300 feet fibre patch lead at 10gig , if they don't want to wrap the lightpeak in copper and push power down that, then il make do with antiquated RJ45 POE (at least that's one good thing) and tape that to the 100 meter fibre and run the remote device that way if needs be.

OC the reason he says three meters for his USB3 is the simple fact that YOU CANT Even Run and power even a simple Wireless 11g/N stick over a that same USB extension on Most USB PC ports Today, if they do make a PO Lightpeak copper wrapped Fibre patch lead then you may actuall use that for remote router to Lightpeak Backbone you might set yo in your home etc...

who's this engineer trying to kid ?, the FAR Higher 10Gig/s data rate is what matter to end users TODAY, people will always find ways to power kit if need be.

and a simple generic 10 gig 4 port Ethernet Over lightpeak on a PCiE card with 4 fibre patch leads in a self install LAN/WAN kit for $99 is going to fly off the retail shelves in early 2011 if they make them.

its just a shame we cant buy them RIGHT NOW, as i/most end users have lots of Christmas holidays money/credit burning a hope right now just waiting to be spent on something worth while like light peak LAN Kits, all the other potential use stuff is just added gravy, Ethernet Over Lightpeak all the way ;)